July 2011
CPSS IOSCO Consultative report on
Principles for Financial market infrastructures
Comments from NASDAQ OMX
The NASDAQ OMX Group, Inc. delivers trading, exchange technology and public company services across six continents, with more than 3,600 listed companies. NASDAQ OMX offers multiple capital raising solutions to companies around the globe, including its U.S. listings market, NASDAQ OMX Nordic, NASDAQ OMX Baltic, NASDAQ OMX First North, and the U.S. 144A sector. The company offers trading across multiple asset classes including equities, derivatives, debt, commodities, structured products and exchange-traded funds. NASDAQ OMX technology supports the operations of over 70 exchanges, clearing organizations and central securities depositories in more than 50 countries. NASDAQ OMX Nordic and NASDAQ OMX Baltic are not legal entities but describe the common offering from NASDAQ OMX exchanges in Helsinki, Copenhagen, Stockholm, Iceland, Tallinn, Riga, and Vilnius.
NASDAQ OMX welcomes the opportunity to comment on the Consultative report on Principles for market infrastructures which is published at a very important time.
NASDAQ OMX provides comments primarily from the perspective of operator of derivatives CCPs, for equity, fixed income and commodity derivatives. We however also own CSDs in the Baltic region, so general comments are also applicable to FMIs in the form of CSDs. Furthermore, we take a general perspective of how to best regulate and organise the market in the interest of achieving efficient, stable and well-functioning financial markets.
NASDAQ OMX is a member of the European Association for CCP Clearing Houses (EACH) and of CCP12. Below we develop our own comments in addition to those of EACH and CCP12.
General comments
NASDAQ OMX supports the development of appropriate standards for market infrastructures (FMIs) at a global level. As FMIs are becoming increasingly important in the aftermaths of the financial crisis - especially CCPs are seen as a means to reduce risks1 - a global approach is crucial. Going forward, the global regulatory approach will become increasingly important, with continuous progress of globalization of business.
We urge CPSS IOSCO – as well as other regulatory bodies – to ensure appropriate coordination among the various regulatory frameworks that exist and are being developed, especially for CCPs. As far as possible, inconsistencies in rules and principles must be avoided. The obvious regulatory frameworks to point at are, besides CPSS IOSCO Standards, Basel III, EMIR, ESCB/CESR Recommendations and Dodd-Frank. Furthermore, we note that adding to the fragmented regulatory environment is the fact that the process for being classified as a Basel III qualifying CCP is not clarified.
Especially regarding the classification of whether or not a CCP is Basel III compliant, we have the following comments. If EMIR is more demanding demanding/restrictive, there will be an uneven playing field among European and non-European CCPs. If CPSS/IOSCO is more demanding – CCPs would have to comply with the higher CPSS/IOSCO standards to be a qualified CCP. Examples where CPSS/ IOSCO appears more restrictive (depending on the final outcome of the EMIR wording):
(i) Letters of credit/on-demand guarantees as collateral. Under the current proposal of EMIR, bank guarantees may be permitted as approved collateral from non-financial counterparties. Under the CPSS/IOSCO draft however, bank guarantees appear to be acceptable “in unique circumstances and subject to regulatory approval, a guarantee fully backed by collateral that is realizable on a same-day basis may serve as collateral”. See more of our comments under Principle 5.
(ii) Exposures to be covered by clearing capital and default fund in EMIR – two largest exposures. In CPSS/IOSCO it is not determined – one, two or one +.
(iii) Requirements on segregation of customer assets and positions are more detailed in EMIR than CPSS/IOSCO.
Also, maintaining a level playing field requires efforts beyond developing appropriately harmonized rules. Supervisory authorities also need to ensure that the application of the Principles support fair and sound competition. We appreciate the work done so far by IOSCO in developing cooperation among supervisors and encourage further progress.
Global as well as regional and local regulatory frameworks nevertheless need to allow room for flexibility, as several equally acceptable models for FMIs exist and should continue to exist. The highest standards for risk management and governance are fundamental. Still, there are several ways of how to achieve this. In order to not hamper sound development of efficiency and innovation, regulation should refrain from applying the one-size-fits-all approach. Also, local legal environment, market structure and other conditions may motivate different solutions. We support the approach that the overarching global framework should include principles and state the goals, while local rules may display different means of achieving the same principles and goals. One must also remember that FMIs come in many different sizes. Due consideration must be taken to the fact that smaller players should not be forced to adapt their models to those of the bigger ones simply in the name of
harmonization. On the contrary, the smaller ones play an important role in the competitive landscape and their models may play an important role. It goes without saying however, that all FMIs of all sizes must live up to the necessary regulatory standards.
Lastly, some of the recommendations, especially under principle 7, will require significant changes to the organisation and business conditions, especially for CCPs. Sufficient time is necessary for FMIs to adapt to new recommendations. 12 months may be too short.
Below we provide specific comments to the individual principles. Comments on a selection of the principles
Principle 2: Governance
We agree with the Principle 2 and also with what is expressed in the Key considerations. Good governance is a precondition for sound management of a company and for creating shareholder value. However, we have some concerns regarding the details in the Explanatory note.
We hesitate on the link between good governance and incentives for the board to support the public interest (3.2.1). We share the view that good governance supports the incentives for the board to pursue the best interests for its shareholders and other stakeholders. Good governance also facilitates effective monitoring, for instance by supervisors, which indeed may take place in the public interest. This is however not the same thing as placing on the board an obligation to always pursue the public interest. This must be the purpose of other regulatory frameworks and of supervisors. We believe our views are in line with the OECD Principles of Corporate Governance.2
The governance of any company or company group is overall affected by how the group is organized. Any detailed recommendations on how to set up the details of the governance structure must
acknowledge the overall organisation of the company or company group and applicable local legislation. This is to some extent recognized in paragraphs 3.2.4 and 3.2.5, but the paragraphs that follow thereafter must be clearly linked to this principle starting point in order to avoid confusion in the application of these recommendations.
What we say above may be exemplified by how the company and governance structure is organized around the derivatives CCP operated by NASDAQ OMX in the Nordics. The CCP is operated under a license from the Swedish FSA. It is operated within a Swedish company and it serves participants in all Nordics and beyond. This same company also holds licenses to operate trading venues (regulated market, MTF). The company is part of a Nordic/Baltic company group, where other companies also hold licenses to operate trading venues and offer other financial services, including CSDs. This Nordic Baltic company group is in turn part of a larger global NASDAQ OMX group. Especially the Nordic governance is very much integrated and there is a two-tier board system within the Nordic part of the group. At the same time, the parent company of the group, The NASDAQ OMX Group, Inc., carries overall responsibility of the whole group. Many reporting lines and much of the committee structure referred to in the Principles is already established on a group basis. The governance of the CCP needs to be organized in a way that best serves the CCP as well as the businesses, operations and risk management in the whole company group, ensuring the most appropriate reporting lines and enabling management and board to take the due considerations and responsibilities. For instance in paragraph 3.2.10, it should be clarified that the application of this recommendation must take into account the overall company structure in which the CCP is operated.
The relationship between the board risk committee that a CCP is expected to have (3.2.12) on the one hand, and the rules on a risk committee that are being developed within EMIR (tentatively article 26) on the other hand, is unclear. One risk committee should be sufficient. Whether or not a CCP board should be „expected‟ to have a separate risk committee may depend on the more detailed local
legislation and there may be room for more than one way of organizing risk committees. Again, this is recognised in 3.2.4 generally for governance and the recommendations regarding the risk committee should be understood in the light of these general statements.
The recommendation on model validation (3.2.13) should be clarified. The text in the Report is inconsistent in the use of terms such as independently and qualified and independent party. It is unclear if the recommendation is internal or external validation. Validation is best done within the FMI, where the expertise is. This would result in better quality validations than if the validations were done by external parties, albeit independently in the sense that it would be done by a third party. We agree internal validation should be done by a function that is separated from the actual function it validates. External validation may still be done when the FMI or its supervisor deems it necessary. Further, regarding validation, we agree an FMI should validate on an ongoing basis its models and methodologies. We however question the added benefit of requiring frequent validation of models when there have been no changes in conditions.
Principle 3: Framework for the comprehensive management of risks
We question the requirement to provide the participants and customers with the capacity to manage and contain their risks as well as the requirement to regularly review the material risks that are posed to other entities. Each business is solely responsible for its own risk management. It is difficult in practice for any FMI to assess the risks it poses to other entities. Understanding the risks posed to each individual customer would require an FMI to gather a considerable amount of information to which it does not have access today. Fulfilling this requirement would generate considerable costs and complexity for FMIs.
Especially regarding a CCP, it even offers risk reduction to its counterparties as its main business. Still, a CCP cannot be required to provide to another business the capacity to manage risks, it cannot take on such liability.
We suggest the recommendation instead includes a requirement to provide the necessary information (subject to further specification) for participants and customers to be capable of managing their risks.
Principle 4: Credit risk
We fully support that an FMI should perform necessary testing of stress situations. This must include both backtesting and testing of potential future scenarios. This is already done today and is necessary especially in order for a CCP to manage credit risks under potentially extreme conditions.
We however question the relevance of performing reversed stress tests beyond what is considered „extreme but plausible‟. The reverse stress test for a CCP aims to define under which market conditions a CCP‟s resources will be insufficient. It must be remembered that a CCP‟s resources consist of margin for the daily „normal‟ conditions, whereas other resources (default fund and the CCPs own funds) are used to cover potential future exposures under extreme, but plausible,
conditions. And, as the size of the default fund is recalculated on a daily basis and the CCP can ask for more contributions at all times (normally quarterly but more often if deemed necessary) the resources of the CCP will change as the market becomes more volatile. In more volatile periods a CCP may also increase the risk parameters which will give higher margins. The stress-test and the assumption used are already designed to ensure that the CCP at all times have sufficient resources. It is of course always possible to model “even worse” scenarios (like a default of all members) but the additional value of such a test is very limited.
Key Consideration 5 refers to a daily backtest. However, backtesting is based on a large number of observations during a long history, so the result cannot be expected to change significantly from one day to another (unless in the case of an extreme market situation). It seems more appropriate to conduct backtests on a less frequent basis, except under extreme market conditions. Furthermore, backtesting should be performed on participant level rather than at the level of an individual product or spread, since the exposure of the CCP is to the defaulter‟s overall portfolio.
We also believe the link between stress tests and initial margin gives an incorrect picture. Margin is calculated by a CCP with the objective of establishing the level of risk on a position under normal market circumstances, i.e. within a pre-defined confidence interval. Stress testing is performed in order to estimate the potential loss on that same position under extreme but plausible stress scenarios with the aim of establishing which amount of potential stress loss should be covered over and above the margin available. In practice, this potential stress loss will be covered by a default fund or other financial resources of the CCP in addition to the initial margin. Margin and stress testing therefore both cover different levels of risk and should result in complementary risk buffers.
Principle 5: Collateral
No CCP should be forced to accept a certain type of collateral. It must be up to the CCP itself to decide whether or not to accept a certain type of collateral. This is a fundamental principle. The range of collateral that may be acceptable should not be overly restrictive. Some types of collateral may require extremely conservative haircuts, which does not make them unsuitable as collateral per se. Especially when collateral is posted under a pledge arrangement as opposed to a title transfer. A broad range of acceptable collateral– always with appropriate and conservative haircuts – provides the financial system with a healthy dose of diversification and it allows for ensuring the best collateralization for each type of participant, client, position and market conditions at each time. For instance, for a sold call option the absolute best collateral is the underlying share, not a AAA-rated government bond or cash. And similarly, instead of forbidding a type of collateral that may potentially pose a wrong-way risk, such collateral should still be accepted, but the CCP must be required to closely monitor this collateral and have procedures in place for managing such wrong-way risks. As a general comment, a CCP does not always need immediate access to collateral. This is because of natural „delays‟ in the settlement. For forwards, the daily settlements are low compared to settlements in futures. A close out of the positions does not change the settlement cycle for forwards, i.e. the CCP could wait two to three days to have access to the collateral and still be able to timely fulfill the settlement obligation, because of these natural „delays‟ in the settlement cycle.
Letters of Credit / on-demand guarantees should be an acceptable form of collateral
We have specific and strong views in respect of Letters of Credits or on-demand guarantees (hereinafter on-demand guarantees) as collateral – we believe they must be recognized as a type of collateral that CCPs may opt to accept. On-demand guarantees can be an equally sufficient type of collateral as for instance highly liquid financial instruments. Such guarantees must however fulfill certain criteria, most importantly being available on-demand (making them „highly liquid‟ in
practice). The accepting of on-demand guarantees as collateral must also be subject to a consideration of other risks, such as concentration risk.
On-demand guarantees are an accepted form of collateral within the energy commodity clearing space world-wide; for example being used in North America by the Natural Gas Exchange as well as our own clearing affiliate NOCC. In Europe, on-demand guarantees are frequently used as collateral in the NASDAQ OMX CCP for commodity derivatives.
The use of on-demand guarantees in the energy clearing space is mainly related to the characteristics of the end-users participating in these markets. Electricity producers, energy exploration and energy distribution companies do not possess the same levels of readily pledge-able liquid assets as do the main users of a financial derivatives CCP, which tend to exhibit a member base more heavily
weighted towards financial institutions. Consequently, on-demand guarantees have become normally used as collateral for the energy markets. Furthermore, large electricity producers, which often rely heavily on this type of collateral, are considered a “right-way risk” to the CCP from a counterparty exposure perspective, as they hedge their commercial risk towards the CCP via forward contracts, and therefore have receivables that offset all of their trades. A general prohibition of this form of collateral could thus have severe negative consequences for the energy commodities industry. Given the large size of their commercial hedging activity, large utility companies in Europe‟s most liquid power market - the Nordic region - provide vital liquidity to the overall market.
Exchange traded electricity forward contracts have different settlement schedules. Forwards typically have up to several months or years until settlement. In case of an on-demand guarantee bank default, where the collateral is backing a portfolio of forward contracts, the CCP does not need the liquidity until settlement, which could be years later. In those circumstances the issue with “on-demand bank guarantees not being immediate accessible” compared to other assets such as cash, would be less important because under such circumstances the CCP will require immediate replacement of the failed provider with a new viable (credit worthy) one. The main risk is if the utility company and the
guarantor default simultaneously. However due to the longer settlement cycle, the CCP should have ample time to mitigate such a liability. In the case of futures however, the contracts would be settled more frequently. Liquidity requirement should then be assessed and handled separately (i.e. with sufficient liquidity arrangement set up by the CCP).
The liquidity profile of an on-demand guarantee is quite strong as long as the bank maintains high short-term rating and is subject to rigorous regulatory capital requirements for contingent obligations. Moreover, the value of the on-demand guarantee is stable and does not fluctuate with the market volatility from day to day as securities do. On-demand guarantees introduce no more credit or liquidity risk to the CCP than taking pledged cash and sweeping it directly into a bank‟s demand deposit account or making a direct investment in a bank‟s short-term commercial paper program.
Over-constraining collateral may have the unintended consequence of pushing trading to avoid the transparent wholesale markets and structures, such as exchanges and CCPs. Instead, market participants may revert to bilateral trading of physical contracts taking place OTC, with less transparency and without clearing. Not accepting on-demand guarantees is likely to dramatically increase the cost of clearing as energy market participants would have to provide other types of collateral not readily available to them. In turn, pushing utility companies into more OTC trading would decrease market transparency and increase costs. We believe that in the spirit of transparency in current financial reform, creditworthy non-financial hedgers should have the option to be CCP participants at a reasonable cost, as they currently are in the Nordic electricity markets.
In light of the above arguments, we insist that the CPSS-IOSCO recommendations make it clear that on-demand guarantees are an acceptable form of collateral. The example of on-demand guarantees as collateral also illustrates general points made above, i.e. the need for allowing flexibility, allowing a diversity of business models and adaptation to local practices.
Principle 6: Margin
Regarding price information (3.6.4), a CCP should have the organization and competence to be allowed to perform the valuation of its price data and valuation models by using internal resources. Independent third parties should not be required (see also above under Principle 2).
The wording on initial margin methodology says that the CCP should select an appropriate close-out period for each product cleared. We however believe the CCP should have the possibility to cluster related products, not needing to assess every single product. This should be clarified. Furthermore, again, a CCP should have the organization and competence to be allowed to perform the valuation of its initial margin models by using internal resources.
We reiterate that stress testing of margin coverage under highly stressed hypothetical scenarios may not add significant value, since margins are intended for normal conditions.
Principle 7: Liquidity risk
We strongly support timely settlement and agree that a CCP should have sufficient liquid resources to meet required margin payment. However, same-day close out does not necessarily require same-day liquidity. A close-out or hedge will require settlement in three days for most products, meaning that the pledged collateral could be utilized. For forwards, the timeline is longer than for futures. Possibly, a time dimension should be built in to the requirements regarding liquidity resources. In fact, this requirement for same-day close out or hedging appears to contradict recommendations elsewhere in the report, such as in 3.13.1: (c) limiting disruptions to the market; (e) managing and closing out the defaulting participant‟s positions and liquidating any applicable collateral in a prudent and orderly manner. See also our comments under Principle 5 above.
And again, reverse stress testing (3.7.15) provides very limited additional information in relation to the standard stress tests. It is also unclear how one should build scenarios for conducting reverse stress test in practice or how frequently stress tests should be conducted. See also our comments under Principle 4 above.
There are differences between market/credit risk and liquidity risk. This calls for keeping a distinction between capital requirements for simultaneous default of the two largest participants (market/credit risk) on the one hand, and the liquidity needed for simultaneous default of the two largest exposures (liquidity risk) on the other. For the market/credit risk, “cover two” is appropriate. For the liquidity risk, “cover one” is appropriate. First of all, market/credit risk buffers (margin, default fund) must cover losses resulting from liquidation. Therefore these buffers can be used only once during the default management process. Liquidity buffers, on the contrary, are “self replenishing” in the sense that they are used in the settlement process and freed up again in the liquidation process. A scenario determining the liquidity need by assuming the default of the clearing members posing the two largest liquidity requirements would therefore also have to assume that these defaults take place at exactly the same time and result in liquidity needs at exactly the same time. As mentioned, this is different for market/credit risk since margin and default fund contributions can be used only once to cover the loss. The possibility of an exactly simultaneous second failure giving rise to an additional liquidity need is far less than the (already remote) possibility of a second failure requiring additional default resources before they have been replenished from the first default, some days earlier.
Principle 15: General business risk
The definition of “going concern” should not be modeled around a requirement to hold equity capital equal to 6, 9 or 12 months of expenses. No single model is equally appropriate for all FMIs, based on differences in type of FMI, company structure and set-up. Each FMI should define its own model and measures, which should be approved by the local regulator.
In case a set time period is preferred, 6 months expenses should be sufficient, since this will be the closing down period on average under the assumption that no revenues at all will be generated. Furthermore, depreciations of tangible and intangible assets should be deducted from the expenses, since no additional investment would be conducted in an orderly winding down process. The expense would be directly booked against the asset which is depreciated. Consequently, there is no need to book this expense against asset in terms of cash holdings.
In addition, we note that the European Commission is currently considering the possibility of
imposing capital requirements for CSDs in its future legislation. In this context, it is very important to achieve consistency between the future EU CSD legislation and the CPSS-IOSCO Principles.
Principle 17: Operational risk
The requirement for FMIs to "manage the risks [their] operations pose to other FMIs" is likely to be difficult to implement in practice. As mentioned above under the comments to Principle 3, the need for FMIs to systematically assess the risks they pose to other entities would require them to gather a considerable amount of information to which they do not always have access today.
Principle 19: Tiered participation arrangements
Full risk monitoring of indirect clients appears impossible in practice. The NASDAQ OMX CCP in the Nordics alone has 15 000 indirect client accounts. Also, the identity of the indirect clients is unknown to the CCP in most cases. Each CCP must have the right to define its own rules for how and when an indirect client should be monitored and what action to take in case of increased risk. Further,
it is unclear how this principle could apply to CSDs. Therefore, a clearer distinction between different types of FMIs would be necessary with regard to this principle.
